I used to own a Nex7 with the Sony Zeiss 24mm lens, and the Sigma DP2M. I currently have the Sony RX1 with electronic viewfinder, and also a Nikon D800E.
My favorite focal length in photography is a DSLR full frame equivalent of 35mm so spent a boatload of money on the RX1. The idea of whether it is worth the money is something that I can't even answer for myself right now, but I would like to comment on the other aspects. My experience with the RX1 is very positive image quality wise, and overall I have a very similar impression to that Michael stated in his review of the camera.
I find the RX1 to be far superior in image quality to the Nex7 w/24mm lens - close up and also for landscapes with the lens focused at or near infinity. Superior in sharpness, color, contrast, dynamic range. This conclusion is from pixel peeping the images taken in raw format and processed in LightRoom. There is some very slight purple fringing with the RX1 at f2 to f4 in some high contrast situations, but I think the Nex7 24mm combo had it or more as well. The RX1 lens seem to have a fair amount of vignetting but it's easily corrected if needed of course.
I've tried a bunch of 35mm focal length lenses on the D800E (rentals and some purchase w/returns) - including the Nikon 35mm f1.4G, Zeiss 1.4, Zeiss f2, Nikon AF-D f2, Nokton 40mm, and more recently a Sigma f 1.4 and another round with some of the other lenses. My primary subject is landscape and I concentrated on sharpness at near infinity across the entire image. At the individual pixel level so to speak I am happier with the RX1. It amazes me just how sharp the lens is, even at f2. Much better uniformity at all apertures as compared to most of my experiences with the D800E. I had some commercial prints made on a Lightject (not inkjet) and regardless of capture and post sharpening the differences don't really show up that much at all up to 16x24", in sharpness. Digital LJ or Chromira, the great equalizer for the digital age. Maybe my techniques are deficient and/or maybe I need to move to inkjet printing, but that's another whole topic.
If have found focusing with the D800E to be more problematic than with the RX1. Live view is not great on the 800E and I found extremely tiny rotation differences of the SLR lenses to make a significant difference in sharpness (again, at the pixel peeping level). I also found a wide variance in performance of all the lenses - for example the Zeiss 1.4 was the best for landscapes when stopped down to f8. It was a rental. I bought the lens but it was far inferior to the rental. Tried another copy. Even worse than the 1st. The Nikkor 1.4 apparently had tilted elements and the right side was way off from the left side of the image. Tried renting some same lenses again and the results were worse - either I got lazy in focusing or these samples were beat up. The Sigma was better than all of them with the possible exception of the Zeiss f2 lens, but none of the results were as good as when I first rented lenses a few months ago. The Sigma has gotten great reviews but I don't think any of them have been on the 36mp Nikon. So at the moment I do not have a 35mm lens for the 800E. The RX1 is so good that it mitigates this situation to a great extent.
The RX1 is such a specialized and different tool than a DSLR - much smaller, lighter, and extremely quiet. I don't like electronic finders, but the one for the RX1 is good. I wish Sony had put it in the RX1 body even though it would made the body a little larger. I preferred the Nex style grip much more than the RX1 - I found it so much easier to hold the Nex than the RX1. I don't like the add-on electric finder sticking up above the body of the Rx1. However on the upside the 90 degree rotation ability of that finder can come in very handy for some shots.
I customized the buttons on the RX1 slightly and have found it very easy to shoot with in a variety of conditions. It is much easier to take with me for casual outings than a DSLR. It provides great image quality. It oozes quality in the way it is built and how the controls feel. I've other smaller format cameras including several micro-4/3 but there really is no comparison image-quality wise with the RX1.
I don't have my Sigma DP2M anymore -I found it too time consuming to shoot in raw and process the color and contrast to my satisfaction, but the major reason was the 45mm equivalent is not a favorite focal length of mine. Comparing the DP2M to the RX1 is difficult because of the difference in equivalent focal lengths. Yes the DP2M is much sharper on a pixel basis, but it has a smaller field of view. Cropping the RX1 images to the same field of view as the DP2M does seems to yield a sharpness advantage to the DP2M, but that doesn't do me much good when I want a 35mm type of field of view.
As to bokeh, indeed this must be one of personal preference, and also it does vary from subject to subject, various lighting, distance to subject /foreground/background, etc. For me I find the RX1 bokeh to vary from pretty good to not very good. I like the bokeh from the DP2M (better than the RX1). For the DSLR lenses I liked the two Zeiss lenses better than all the others including the Sigma.
After having gone through many iterations of equipment (with more sure to come) and doing my own comparison, I have reached the conclusion that most tests on the web and in photography magazines, can be helpful but also can be misleading - leaving one in no-man's land. I think this is due to sample variation and variations in methods of testing and shooting test charts that are relatively close to the camera as compared to landscapes. As just one example, you'll see test of lenses on the Sigma from DxOMark, a resolution test on Lensrentals, other web lens test sites and one in the current issue of Pop Photo. The one from Pop photo doesn't give the Sigma the highest marks as the other sites have, yet that test report coincides with part of my experience as noted above - the Zeiss lenses being better than the Sigma, at least the 1st time around with the 1.4. I have found sample images on the web to be almost useless for any purpose - too many variables.
The most useful tests I have found are here on LL by Michael, and some on Diglloyd and on Reid reviews.